Assessment of body temperature is such a routine aspect of initial patient encounters that the accuracy of thermometry is readily taken for granted. Such trust in thermometry has been extended to the new infrared ear or tympanic thermometers.

The infrared sensors of these devices are inserted to variable depths in the ear canal and measure, without direct thermal contact, the infrared radiation emitted by the tympanic membrane or adjacent skin linings. An enclosed cavity delineated by the tympanic membrane, ear canal, and the infrared probe tip is created, with the probe tip as the coldest surface. Water vapor derived from insensible perspiration is also present.

The temperature of the window surface of the inserted infrared sensor is lower than the dew-point temperature of the humid ear canal, leading to rapid condensation of water vapor on the probe window. Because the water condensate is opaque for infrared radiation, the result is a rapid and time-dependent decrease in the infrared signal. The instrument registers a decrease in tympanic or ear canal temperature (Figure 1, curve a) while reflecting the increased energy absorption by the water condensate. This dynamic can be tested by inserting an instrument in the ear and taking repetitive measurements without removing it from the canal, which will result in a time-dependent decrease in temperature readings (Figure 1, curve a). Probe design, its heat capacity and exchange characteristics, its window size and covering, and its actual temperature affect the kinetics and extent of the apparent temperature change. Prewarming the probe tip to a temperature slightly greater than body temperature and repeatedly measuring the tympanic temperature prevents condensation and results in unaffected infrared signals and unaffected temperature readings (Figure 1, curve b).

View larger version (14K): [in this window] [in a new window]   Figure 1. The time-dependent decrease in apparent human infrared (IR) tympanic temperature and its elimination. Curve a was obtained with an instrument probe at 22 °C before insertion into the ear canal. Curve b was obtained with an instrument probe prewarmed to 38 °C before insertion.

Condensation as a potential source of error in infrared thermometry can be further shown by inserting an infrared ear thermometer into the open end of a thermostated (36.5 °C) metal tube. If the tube cavity contains a water droplet, repeated measurements of the temperature of the tube with the infrared thermometer at room temperature show rapid decreases in temperature. When the probe tip is prewarmed to 37 °C, repeated measurements show no decrease in temperature readings. The same is true when the tube contains silica drying gel instead of a water droplet.

Although our findings are provocative, further studies are needed to assess the clinical significance of this potentially serious problem.